Learn the stock market › Studies
What happens after a stock falls 10% in a day? A month and a year later, every fall since 2000
About the same as after an ordinary day. A month after a fall of at least 10%, the stock stood higher in 56% of cases and a year later in 69%, against 58% and 72% from an ordinary day at the same companies. We counted 251 falls at 20 large US companies since 2000. Companies that fell and never came back are missing, so the real picture was worse.
The stock stood higher afterwards
What we counted
"Should I buy stocks when they drop?" is a common question. What you should do is not something we can calculate, but what happened afterwards is. A fall here is a trading day when the stock closed at least 10% below the close the day before. Since 2000 that has happened 350 times at the 20 companies we follow.
Several falls in a row are usually the same event, so a company counts at most once a month, from the first fall. That leaves 251 falls. We measure from the close on the day of the fall and compare with the same company from an ordinary trading day.
These are the largest companies today. A company that fell and never came back is not here, and those are exactly the companies the question is about. So every number after a fall is higher than reality.
A month and a year later
| After | Higher after a fall | Higher from an ordinary day |
|---|---|---|
| The 20 US companies, a month | 56% | 58% |
| The 20 US companies, a year | 69% | 72% |
| Stockholm's 30 companies, a month | 58% | 57% |
| Stockholm's 30 companies, a year | 75% | 68% |
After a month the stock stood higher after 56% of the falls, against 58% from an ordinary day. After a year it was 69% against 72%.
The gap in that share is −1.6 points after a month, with a range of chance of −7.8 to +4.7, and −3.0 after a year, with a range of −13.6 to +7.5. Both include zero, so neither can be told apart from chance.
The average after a month was +4.6% against +2.8%. The difference is +1.7 percentage points and the range of chance −1.4 to +4.8. That includes zero, so it cannot be told apart from chance.
The average after a year was +55.4% against +45.0%. The difference is +10.4 points and the range of chance −12.8 to +33.5. That includes zero, so this cannot be told apart from chance either.
A few very large rises pull the average up. The middle fall was followed by a change of +3.0% after a month and +34.0% after a year. The worst year followed Amazon's fall on Jan 5, 2000: −78.4% a year later.
Did the whole market fall, or just the company?
We say the whole market fell when the middle company on our list fell at least 3% the same day. That was true of 64 of 251 falls, and 35 of those came in 2008, 2009 or 2020.
| After | Higher after a fall | Higher from an ordinary day |
|---|---|---|
| The whole market fell, a month | 65% | 59% |
| The whole market fell, a year | 87% | 74% |
| The company fell alone, a month | 54% | 58% |
| The company fell alone, a year | 62% | 71% |
After a fall on a day the whole market fell, the stock stood higher more often: a year later in 87% of cases, against 74% from an ordinary day and 62% when the company fell alone. Those cases sit in few periods, so they are not many independent events. When the market turned up, nearly every company rose together.
When the company fell alone, the stock stood higher less often than from an ordinary day: 54% against 58% after a month, and 62% against 71% after a year.
The average difference after a year for the lone falls is +4.1 points, with a range of −17.0 to +25.2, so it cannot be told apart from chance.
How often did the stock come back?
Back here means the stock closed at least where it stood the day before the fall. That happened within a month after 38% of the falls and within a year after 85%. The middle case took 37 trading days.
Falls more than a year old where the stock has not come back: 1. Furthest away is Netflix after its fall on Oct 22, 2025, now −48.9% from the close the day before.
Did the falls come on earnings day?
Of the 88 falls after a company's first stored report, 41 came on earnings day. How far a stock usually moves on that day is in the study of earnings days, and other causes are in the lesson on why stocks go up and down. How deep the whole market has fallen is in the study of every crash.
The biggest falls
Every fall counts here, also those close together. The numbers on the right run from the close on the day of the fall.
| Company | Day | The fall | A year later |
|---|---|---|---|
| Apple | 2000-09-29 | −51.9% | −37.3% |
| Netflix | 2004-10-15 | −40.9% | +179.2% |
| Nvidia | 2004-08-06 | −35.2% | +199.6% |
| Netflix | 2022-04-20 | −35.1% | +45.0% |
| Netflix | 2011-10-25 | −34.9% | −22.3% |
| AMD | 2002-10-03 | −32.4% | +227.5% |
| Nvidia | 2002-07-31 | −31.8% | +72.4% |
| Nvidia | 2008-07-03 | −30.7% | −15.5% |
| Eli Lilly | 2000-08-09 | −29.1% | +1.8% |
| Netflix | 2004-07-16 | −28.1% | −28.8% |
Company by company
A company has only a few falls, so a single row can shift a lot by chance. The table is sorted by the number of falls, most first. The last column is how often the same company stood higher a year on from an ordinary day.
| Company | Falls | Higher a year later | Higher from an ordinary day |
|---|---|---|---|
| AMD | 52 | 25 of 49 | 57% |
| Amazon | 28 | 14 of 28 | 80% |
| Nvidia | 28 | 22 of 28 | 73% |
| Netflix | 27 | 18 of 27 | 71% |
| Tesla | 24 | 18 of 22 | 69% |
| Oracle | 13 | 5 of 12 | 72% |
| Palantir | 12 | 8 of 11 | 77% |
| Broadcom | 9 | 6 of 6 | 90% |
| Apple | 8 | 5 of 8 | 81% |
| Meta | 8 | 5 of 8 | 79% |
| JPMorgan Chase | 7 | 7 of 7 | 68% |
| Mastercard | 6 | 6 of 6 | 84% |
| Microsoft | 6 | 5 of 6 | 74% |
| Costco | 5 | 4 of 5 | 85% |
| Eli Lilly | 5 | 5 of 5 | 74% |
| Visa | 4 | 4 of 4 | 88% |
| Alphabet | 3 | 2 of 2 | 81% |
| Walmart | 3 | 3 of 3 | 75% |
| Exxon Mobil | 2 | 2 of 2 | 65% |
| Berkshire Hathaway | 1 | 1 of 1 | 76% |
How we counted
- 20 of the 20 largest US companies on our list, from January 2000 or the company's first price to Nov 6, 2026, and for comparison the 30 companies in OMX Stockholm 30. The first fall came on Jan 5, 2000.
- Companies that went bankrupt, were bought or left the exchange are missing. That lifts the numbers, probably most after a fall.
- Closing prices only. A stock that fell 10% during the day and closed higher does not count. The 10% limit is our own, and a different limit gives different numbers.
- Prices include dividends and splits, so an ex-dividend day does not count as a fall.
- We checked every fall against unadjusted closes on 5 November 2026. All the US falls held. In Stockholm, two were not falls and count as unchanged days: Atlas Copco's SEK 40 share redemption on 18 May 2007 and Nordea's ex-dividend day on 29 March 2019.
- A month is 21 trading days and a year 252. A fall newer than that does not count in that column.
- An ordinary day is every trading day at the same company after the first 200, including days with falls. Each fall weighs the same, so companies with many falls weigh more in the ordinary-day column too.
- The range of chance is 95% around the average difference. Companies count per year, because many fall in the same weeks.
- Report dates are the ones we have stored, at the earliest from about 2014.
- Companies are in the currency the stock trades in. Fees and tax are not counted.
You are welcome to use the numbers
Link here, quote the table or embed the chart on your own page. Name ChartCompass as the source and keep the link.